Caren van Roekel

34 Chapter 2 is problematic in patients with ill-defined tumors (2, 52). Furthermore, nor the acceptable absorbed dose in the healthy liver, neither the effective absorbed tumor-dose are defined yet. Recently, Chiesa et al. have developed a method for activity calculation that does take non-uniformity of the absorbed dose into account. In a cohort of 52 patients with HCC, treated with glass microspheres, they studied lesion and parenchyma absorbed dose at voxel level on 99m Tc-MAA SPECT images. The biodistribution of the 99m Tc-MAA and 90 Y-microspheres was assumed to be identical. The degree of agreement between clinical observations of response and several dosimetric variables was compared. A dosimetric variable that accounts for both non-uniformity of absorbed dose deposition and the dose- rate effect was found to be the best variable to predict response of a lesion. The 50% tumor control probability was different for small (<10 cc) and large (>10cc) lesions, with much higher absorbed doses needed for larger lesions, which weakens the predictive power of planning on lesions. There was a high tolerance of the healthy liver tissue to glass microspheres. The best prediction method was found to be the parenchyma mean dose, with a limit of 75 Gy at a 15% toxicity risk (63). Drawbacks of a voxel-based dosimetry approach can be divided in physical factors, such as attenuation, scatter, noise, and partial- volume effects, and in clinical factors, such as respiration. Corrections for these factors should be implemented in the reconstruction protocol (64). 5. TREATMENT 5.1. Treatment angiography Usually, within two weeks after the scout procedure, the treatment procedure takes place. Similar to the angiography procedure for the scout dose administration, intra-arterial access is obtained via the femoral artery. Before infusion of the microspheres, the vascular tree should be investigated carefully. The exact same injection position as during the scout procedure should be chosen to avoid a different intrahepatic distribution. Wondergem et al. have found a relative difference of >30% in absorbed activity per milliliter in as many as 24 of 68 analyzed segments, due to a difference in injection position of >5 mm (48). It is important to make sure that no new hepaticoenteric collaterals have been recruited in the time interval between the scout and therapy procedures.